Direct Prediction of Steady-State Flow Fields in Meshed Domain with Graph Networks
This work addresses fluid flow prediction for computational fluid dynamics applications, but appears incremental as it builds on existing graph network approaches.
The authors tackled the problem of predicting steady-state flow fields in meshed domains by proposing a graph network architecture that processes mesh-space simulations as graphs, achieving superior performance compared to existing methods.
We propose a model to directly predict the steady-state flow field for a given geometry setup. The setup is an Eulerian representation of the fluid flow as a meshed domain. We introduce a graph network architecture to process the mesh-space simulation as a graph. The benefit of our model is a strong understanding of the global physical system, while being able to explore the local structure. This is essential to perform direct prediction and is thus superior to other existing methods.